| Abstract Title: | Prediction of Retention Indices of Liquid Chromatography Coupled to High-Resolution Mass Spectrometry Pixel by Pixel |
| Presenter Name: | Ms Jiani Hu |
| Co-authors: | Ms Denice van Herwerden Mr Alex Nikolopoulos Mrs Viktoriia Turkina Turkina Dr Saer Samanipour |
| Company/Organisation: | University of Amsterdam |
| Country: | Netherlands |
Abstract Information :
For trend analysis, it is important to compare/align different chromatograms with the same components. For instance, within the environmental analysis, this is crucial since different liquid chromatography high-resolution mass spectrometry (LC-HRMS) methods, columns and systems all have an influence on the retention time (tr). For example, a chromatogram of an environmental sample measured with another method and/or column a week ago cannot be directly compared with the data acquired from the same environmental samples from a month ago. Hence, it is important to build a model which makes it possible to align these datasets. The reason for this is that possible trends of emergence concerning chemicals can be earlier recognized, for example, per and poly-fluoroalkyl groups (PFAS). Therefore, this project aims to build a model where the retention indices (ri) of structural unknown compounds can be automatically predicted. This was performed through interpolation of a data set of pesticide samples analysed with LC-HRMS, where the standard ri of compounds were predicted based on two scales. One of the scales is a homologous alkylamide series1 and the other scale is based on a computationally favourable set of compounds, which is called ri of the University of Athens (UoA)2. This step together with suspect screening was used to pre-process the data. Then, a linear model to predict the standard ri from the tr was built. Next, an uncertainty assessment was carried out by using Leave One Out Cross Validation (LOOCV). After fulfilling the uncertainty assessment, the pixel-by-pixel model was built based on the correlation between ri and tr and validated with pesticide samples. With this model, a LC chromatogram can be mapped to ri based on known standards present in the sample. Furthermore, based on this model different chromatograms could be aligned which can be used to find patterns in, for example, environmental samples. With this emergence concerning chemicals could be identified, or earlier actions could be undertaken to minimize the harm to humans and the world. References: (1) Hall, L. M.; Hill, D. W.; Bugden, K.; Cawley, S.; Hall, L. H.; Chen, M.-H.; Grant, D. F. Development of a Reverse Phase HPLC Retention Index Model for Nontargeted Metabolomics Using Synthetic Compounds. J. Chem. Inf. Model. 2018, 58 (3), 591–604. https://doi.org/10.1021/acs.jcim.7b00496. (2) Aalizadeh, R.; Nikolopoulou, V.; Thomaidis, N. S. Development of Liquid Chromatographic Retention Index Based on Cocamide Diethanolamine Homologous Series (C( n )-DEA). Anal. Chem. 2022, 94 (46), 15987–15996. https://doi.org/10.1021/acs.analchem.2c02893.
